Wednesday, March 27, 2019

Get more help using the T2DKP: videos and webinars

Although we try hard to make all of the contents and interfaces of the Type 2 Diabetes Knowledge Portal clear and user-friendly, they can still be difficult to understand—especially for scientists who are not experts in human genetics. It can be hard to know where to get started among the dozens of complex datasets, several data types, multiple phenotypes, and custom analysis tools included in the Portal. So we're starting to produce two different kinds of video content that will complement our written documentation and satisfy all you auditory learners out there.

First, we're creating short videos (5 minutes or less) that focus on particular aspects of the data and features. Our first one (view on YouTube) is an overall introduction to the T2DKP and the Knowledge Portal architecture. Stay tuned for more in the coming months! And if you would like to see a video on a particular subject, please let us know.

Second, we're planning regular webinars that you can join online. The first of these hour-long sessions gave an overview of the project that created the T2DKP, the data it contains, the major entry points to the results, and a preview of future directions. You can view a recording of the webinar here and download the slides here. If you would like to be notified in advance of these sessions, be sure to sign up for our email list.

Please check out these videos and let us know what you think!

Tuesday, March 19, 2019

Exciting new exome collection now in the T2DKP, along with new datasets and features

The latest release of the Type 2 Diabetes Knowledge Portal includes seven new or updated datasets and several new features.

The highlight of this release is the largest collection of disease-specific exomes to date: exome sequences from 20,791 T2D cases and 24,440 controls, from the AMP T2D-GENES collaboration (Flannick et al., 2019). As well as generating T2D association statistics for individual variants, the analysis produced gene-level T2D association scores that can help prioritize research into new drug targets. These associations are available on a new page of the T2DKP, accessible from the home page. Select "View gene associations":

to view a table listing gene-level association scores:

The table may be filtered by using any of 7 different filters or two aggregation tests. Full details of the filters and methods are described in the preprint. Individual-level data from the AMP T2D-GENES exome sequence analysis dataset are also available for secure custom analysis in the Custom burden test (accessible on Gene pages) and Genetic Association Interactive Tool (GAIT; available on Variant pages).

Many other new and updated datasets are included in this release:

  • We've done further analysis of the BioMe AMP T2D GWAS dataset to generate associations for three complications of T2D: chronic kidney disease in diabetics, end-stage renal disease in diabetics, and diabetic neuropathy. 
  • The FinnMetSeq exome sequence analysis dataset (Locke et al., 2019) includes associations for more than 60 phenotypes, many of them new to the T2DKP, from nearly 20,000 samples.
 is a meta-analysis of T2D associations in over 23,000 individuals of African American ancestry (Ng et al., 2014).
  • VATGen GWAS (Chu et al., 2017), which includes associations for several fat distribution phenotypes, has been updated with sex-stratified cohorts
Liver function GWAS (Chambers et al., 2011) analyzed more than 61,000 samples for associations with four liver enzymes.
  • Genetic Factors for Osteoporosis Consortium GWAS
 (Morris et al., 2019) includes associations for estimated bone mineral density and fracture for more than 426,000 UK Biobank participants.

Find more information about each of these datasets on the T2DKP Data page. Results from each of these studies may be viewed in Interactive Manhattan plots and on Gene and Variant pages, and may be searched using the Variant Finder tool.

This release also includes several new features for the T2DKP. In addition to the new page of gene-level results for the AMP T2D-GENES exome sequence analysis dataset, described above, we've added:

  • Bottom-line p-values for variant associations across different phenotypes, available as an additional option in the PheWAS plot on the Variant page. These associations are derived from a new analysis method, based on METAL, that meta-analyzes all datasets in the T2DKP while taking into account the overlaps between sample sets.

A new Analysis modules page
, linked from the T2DKP home page, that provides quick access to tools. Currently, the tools available from this page are the Interactive Manhattan Plot, the Variant Finder, and the Genetic Risk Score tool; more will be added in the near future.

In response to feedback from T2DKP users, we've changed the Common variants tab on the Gene page (which only displayed variants whose minor allele frequency was greater than 5%) to a Top variants
 tab, showing the most significantly associated variants regardless of MAF.

Please contact us with any questions or comments about this new release!

Friday, March 15, 2019

New funding opportunity from FNIH

The Foundation for the National Institutes of Health is announcing a new Request for Proposals (RFP) to augment and expand the Type 2 Diabetes Knowledge Portal.

RFP16, due April 15, aims to facilitate prioritization of potential causal variants or genes in risk loci or credible sets by:

a. Developing preliminary functional annotation data to support strength of evidence for candidate genes for diabetes and complications (cardiovascular, obesity, diabetic and chronic kidney disease, NASH and liver function, diabetic retinopathy, glycemic traits); and/or

b. Generating mechanistic hypothesis for causal genes from cell-based genome-wide functional or perturbation studies (e.g. RNA-Seq, ATAC-Seq, CHIP-Seq, others); and/or

c. Developing and employing novel tools such as CRISPR tiling to combine with readout such as RNA-Seq to enable prioritizing potential causal variants or genes.

Find full details and contact information here.

Wednesday, March 13, 2019

Upcoming T2DKP webinar March 21

Do you need a little help navigating the tools and interfaces of the Type 2 Diabetes Knowledge Portal or other portals in the Knowledge Portal Network? Attend our webinar on Thursday, March 21 at noon EDT! We'll introduce the motivation for creating the Portals, cover the basics of using them to accelerate your research on complex diseases, give you a preview of what's coming next, and answer any questions you have. Join the webinar here:

Dial by your location
        +1 646 876 9923 US (New York)
        +1 646 558 8656 US (New York)
Meeting ID: 469 259 629
Find your local number:

If you're in the Boston area and would like to attend in person (and join us for lunch), please let us know here.

Friday, March 1, 2019

Faster access to tools from the T2DKP home page

We've rearranged some of the links on the Type 2 Diabetes Knowledge Portal home page, as a first step towards offering a central location for analysis tools. The previous link to the Variant Finder tool has been replaced by a link to the new Analysis modules page:

The new page, shown below, offers access to three analysis tools.

  • The Interactive Manhattan plot allows you to choose a phenotype and view variant associations across the genome for that phenotype.  We've added phenotype selection options to both the Analysis modules page and the Manhattan plot page, making it easier to switch your view between phenotypes.  The default view on the Manhattan plot page shows the largest dataset for a phenotype, but when multiple datasets exist, you can select any one to display. For many datasets, LD clumping is available at several r2 thresholds. Clumping reduces redundancy due to association signals from linked variants, pinpointing the most strongly associated variant in a group.
  • The Variant Finder is a versatile tool that allows you to set multiple criteria (phenotype, p-value, size and direction of effect, and more) and retrieve the set of variants meeting those criteria.
  • The Genetic Risk Score (GRS) module takes a set of 243 variants associated with T2D at genome-wide significance or better, as determined by Mahajan et al. in their DIAMANTE (European) study (Mahajan et al., Nat Genet. 2018 Nov;50(11):1505-1513), allows users to select a phenotype and dataset, and computes the p-value for association of the variant set with the selected phenotype. The interface offers two customizing options: the ability to edit the variants in the set, and the ability to filter the samples by multiple phenotypic criteria before running the analysis. The GRS module, which will be augmented and improved in the future, can potentially reveal relationships between phenotypes. 
The new Analysis Modules page will be the central access point for new analysis tools as they are developed, so check back often for updates!